Lunter group

We are interested in understanding how genomes evolve through mutations and evolutionary pressures, and use this understanding to inform research ranging from disease to human ancestry.

Research overview

HLA heatmap
(Click to enlarge) Stampy and Platypus are able accurately to call variants even in highly diverse regions such as the human HLA. Shown are alignment scores representing the similarity of inferred Platypus haplotypes from 1000 Genomes sample NA12878, to known HLA types. Redder hues represent higher similarity.

The field of genomics has been transformed by next-generation sequencing.  This new data is both posing technical challenges and opening up new research questions.  We believe that a thorough understanding of next-generation sequencing data is crucial to its successful application to scientific questions.  

In addition, the wealth of data enables analyses that only recently were inconceivable, addressing both applied clinical and fundamental evolutionary biology questions.  Often these analyses require novel and efficient statistical algorithms. Our group specialises in developing such methods, drawing on the group's expertise in statistics, algorithmics, mathematics, and genetics.

Our interests span a fairly broad range.  In recent years we devoted much effort to the initial processing of short-read data. This has resulted in the development of Stampy, an accurate and widely-used read mapper that forms a key part of the Centre's sequence analysis pipeline.  A good deal of effort is devoted to the development of our variant caller Platypus. This work led to a collaboration with Prof Nazneen Rahman at the Institute of Cancer Research in London, in which we aim to achieve the highest possible level of accuracy in a novel gene panel that is informative about cancer risk.

Other interests include inference of demographic history from whole-genome data.  In thie project we developed a new approximate but highly accurate model to simulate the Ancestral Recombination Graph, encoding the ancestry of a population sample along the genome.  This forms the basis of an inference engine based on particle filters, a method that recently became popular in financial modeling.  We are the first to attempt using this technique on a computational biology problem, with promising results.

We also have some clinically-oriented projects, including one aiming to develop a wet-lab method to efficiently screen for novel transposable element insertions.  It is known that this class of mutations contributes to human disease, but its importance is unknown due to the lack of an efficient screen.  In another project we collaborate with the Oxford Vaccine Group led by Dr Dominic Kelly to characterize the spectrum of B cell receptor molecules before and after vaccination. We hope that a sequence-based approach can help assess the efficacy of vaccines, and may directly identify the antibodies that confer protection against the disease.